All of these are heuristics applied by deployment planner and host/storagepool allocators to deice the order in which resource(pods,clusters,hosts,storage pools) will be considered for VM deployment.
random: This just shuffles the list of clusters/hosts/pools that is returned by the DB lookup. Random does not mean round-robin - So if you are looking for a new host being picked up on every deployment - that may not happen. firstfit: This makes sure that clusters are ordered by available capacity and first hosts/pools having enough capacity is chosen within a cluster. userdispersing: For a given account, this makes sure that VM's are dispersed - so clusters/hosts with minimum number of running VM's for that account are chosen first. Storage Pool with minimum number of Ready storage pools for the account is chosen first. Userconcentratedpod_random: Always choose the pod/cluster with max. number of VMs for the account - concentrating VM's at one pod. Hosts and StoragePools are chosen randomly. Userconcentratedpod_firstfit: Always choose the pod/cluster with max. number of VMs for the account - concentrating VM's at one pod. Hosts and StoragePools are chosen by firstfit. Hope this helps. -Prachi -----Original Message----- From: Caleb Call [mailto:calebc...@me.com] Sent: Monday, October 22, 2012 10:00 PM To: cloudstack-users@incubator.apache.org Subject: Re: About Allocator algorithm of creating VM on Host Can anyone give a definition of each of the models? I have noticed that my VMs always create on a particular node in the cluster even though I have 6 other nodes that are identical in specs to that one. I have tried first fit (makes sense it would do to the same one till it was full), random (I would expect it to NOT always go to the same one) and userdispersing (not sure what to expect with this one, but tried it anyways). In the logs when it's trying to figure out which node to use, it always finds all the nodes and declares them all fit, but it always puts them on the same node. It seems the algorithms don't work as well as they should. I did restart after each change and could see it was using the new method. On Oct 22, 2012, at 4:28 AM, Tamas Monos <tam...@veber.co.uk> wrote: > Hi, > > The 3.0.2 support the following VM allocation algorithms: > > 'random', 'firstfit', 'userdispersing', 'userconcentratedpod_random', > 'userconcentratedpod_firstfit' > > You can configure this in the global configuration options. The default is > random I guess. > I don't think CloudStack will detect any of those you require however I think > hypervisors should. > Your hypervisor cluster (I use ESX) will detect issues and send alerts, I'm > sure there is something for xen/kvm as well. > > Regards > > Tamas Monos DDI > +44(0)2034687012 > Chief Technical Office > +44(0)2034687000 > Veber: The Hosting Specialists Fax +44(0)871 522 7057 > http://www.veber.co.uk > > Follow us on Twitter: www.twitter.com/veberhost Follow us on Facebook: > www.facebook.com/veberhost > > -----Original Message----- > From: Lucy [mailto:no1l...@gmail.com] > Sent: 22 October 2012 10:45 > To: cloudstack-users@incubator.apache.org > Subject: About Allocatoralgorithm of creating VM on Host > > Dear all, > > > > > > I have a question about allocator algorithm of creating VMs on Host. > > > > What's the default allocator algorithm to allocate VM on Host in Cloudstack? > > > > And are there any other choice? > > > > Can Cloudstack detect heavy loadbalance ,for example one host's CPU/memory is > greater than 80%? > > > > Thanks, > > Lucy > > > >
